Calculate The Attributable Fraction In Exposed Cases

Epidemiology Calculator

Use this clinical and public health calculator to estimate the proportion of cases among exposed individuals that are attributable to the exposure. Choose your preferred input method below.

Expert Guide: How to Calculate the Attributable Fraction in Exposed Cases

The attributable fraction in exposed cases, often written as AFe (also called attributable proportion among the exposed), answers a highly practical question in epidemiology: among people who were exposed and developed disease, what proportion of those cases can be attributed to the exposure itself? This metric is essential in environmental health, occupational medicine, pharmacoepidemiology, infectious disease, injury prevention, and chronic disease control. It helps translate statistical associations into impact-focused language that clinicians, policymakers, and risk managers can use.

If you are evaluating a harmful exposure such as tobacco smoke, asbestos, uncontrolled hypertension, excess alcohol use, or occupational silica dust, AFe provides an estimate of preventable burden among exposed individuals. In practical terms, if AFe is 0.60, then about 60% of disease cases among exposed people are attributable to the exposure and could theoretically be avoided if that exposure were removed and all else remained equal.

Core Formula and Interpretation

The most common expression uses relative risk:

• AFe = (RR – 1) / RR

Equivalent incidence-based form:

• AFe = (Ie – Iu) / Ie

Where Ie is incidence among exposed, and Iu is incidence among unexposed. These formulas are algebraically equivalent when RR = Ie / Iu. AFe ranges from negative values to nearly 1. In most harmful exposure contexts, you expect AFe between 0 and 1. An AFe near 0 suggests little excess disease among exposed. An AFe near 1 indicates that most disease in exposed cases is attributable to the exposure.

Step by Step Calculation Workflow

1. Define exposure and outcome precisely (for example, daily smoking and incident lung cancer over 10 years).
2. Select a valid effect estimate from a suitable design (cohort RR preferred when available).
3. Check risk measure compatibility. If you only have OR, consider converting to RR when outcome is not rare.
4. Compute AFe using RR or incidence formula.
5. Convert to percentage if needed: AFe × 100.
6. Interpret causally only if confounding, selection bias, and measurement error are addressed.

Worked Example

Assume a cohort study reports RR = 2.5 for disease in exposed versus unexposed. Then:

• AFe = (2.5 – 1) / 2.5 = 1.5 / 2.5 = 0.60
• Interpretation: 60% of disease cases among exposed individuals are attributable to the exposure, under the assumptions of the model and study design.

Incidence version example: Ie = 0.24 and Iu = 0.08. Then AFe = (0.24 – 0.08) / 0.24 = 0.16 / 0.24 = 0.667. So about 66.7% of cases among exposed are attributable to exposure.

When You Only Have Odds Ratios

Case-control studies often report odds ratios, not risk ratios. For rare outcomes, OR approximates RR well. For common outcomes, direct OR-based AFe can overstate impact if interpreted as risk-based causation. A practical solution is to convert OR to RR using baseline risk in unexposed:

• RR = OR / ((1 – P0) + (P0 × OR))

Then compute AFe from RR. This calculator supports that conversion path so your estimate remains interpretable for risk contexts.

Comparison Table: Example Relative Risks and Estimated AFe

Exposure and Outcome	Approximate Effect Estimate	Estimated AFe	Interpretation
Cigarette smoking and lung cancer (current smokers vs never smokers)	RR about 20.0	(20 – 1)/20 = 0.95	About 95% of lung cancer cases among current smokers attributable to smoking.
Hypertension and stroke	RR about 4.0	(4 – 1)/4 = 0.75	About 75% of stroke cases among hypertensive individuals attributable to elevated blood pressure.
Obesity (severe) and type 2 diabetes	RR about 7.0	(7 – 1)/7 = 0.857	Roughly 85.7% of diabetes cases among severely obese individuals attributable to that exposure state.
Asbestos exposure and lung cancer	RR about 3.0	(3 – 1)/3 = 0.667	About two-thirds of lung cancer cases among asbestos-exposed groups attributable to asbestos exposure.

Estimates above are rounded teaching examples based on widely reported epidemiologic ranges from major public health literature. Exact values vary by population, intensity, duration, and adjustment set.

Comparison Table: How Different Inputs Change AFe

Scenario	Ie	Iu	RR	AFe
Low excess risk setting	0.06	0.04	1.5	0.333
Moderate excess risk setting	0.12	0.06	2.0	0.500
High excess risk setting	0.28	0.07	4.0	0.750
Very high excess risk setting	0.50	0.05	10.0	0.900

What AFe Is and Is Not

• AFe is exposure-specific among exposed persons. It does not represent total population burden.
• AFe is not the same as population attributable fraction (PAF). PAF depends on exposure prevalence in the whole population.
• AFe is not proof of causality by itself. It assumes the association reflects causal effect after bias control.
• AFe depends on effect measure quality. Poor confounder control gives misleading fractions.

Frequent Mistakes and How to Avoid Them

1. Using OR as RR when outcome is common. Fix this by converting OR to RR using baseline risk.
2. Mixing adjusted and unadjusted estimates across steps. Use a consistent, adjusted measure whenever possible.
3. Ignoring reverse causality and time-order errors. Ensure exposure precedes outcome.
4. Forgetting uncertainty. Report confidence intervals for RR and propagate uncertainty to AFe when possible.
5. Confusing relative and absolute impact. Pair AFe with absolute risk difference for clearer communication.

Advanced Interpretation in Clinical and Policy Contexts

AFe helps individual-level counseling and subgroup intervention planning. For clinicians, it can support risk communication in high-exposure patients. For occupational health teams, it helps justify engineering controls and personal protection investments. For environmental regulators, AFe informs where burden reduction among exposed groups may be largest. In health systems, AFe can prioritize targeted screening or prevention in high-risk exposed cohorts.

However, high AFe does not automatically indicate high total burden. A rare exposure can have very high AFe but still account for relatively few population cases. Conversely, a modest AFe combined with a very common exposure can generate substantial population-level burden. That is why AFe should be interpreted alongside exposure prevalence, absolute risk, and feasibility of intervention.

Authoritative Resources for Deeper Methods

• CDC Surgeon General Reports on smoking and disease risk
• NHLBI Framingham Heart Study resources
• Boston University School of Public Health epidemiology teaching module on attributable risk

Practical Reporting Template

A concise way to report findings is: “Among exposed individuals, the attributable fraction was X% (derived from RR = Y). This suggests that approximately X out of every 100 observed cases in the exposed group are attributable to exposure, assuming model validity and adequate control of confounding.” For technical reports, also include data source, design, adjustment covariates, follow-up period, and sensitivity analysis assumptions.

Used correctly, attributable fraction in exposed cases is a high-value metric that bridges epidemiologic evidence and practical prevention decisions. The calculator above is designed for transparent, reproducible use with RR, incidence, or OR-based conversion pathways.